Speech Recognition Method for Selecting a Combination of List Elements via a Speech Input

• US 20100305947A1
• Filed: 06/02/2010
• Published: 12/02/2010
• Est. Priority Date: 06/02/2009
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Schwarz, Markus, Haiber, Udo, Gerl, Franz, Schulz, Matthias, Biedert, Marc, Hillebrecht, Christian
• Granted Patent: US 8,666,743 B2
• Time in Patent Office: Days
• US Class Current: 704/252
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning

Abstract

The invention provides a speech recognition method for selecting a combination of list elements via a speech input, wherein a first list element of the combination is part of a first set of list elements and a second list element of the combination is part of a second set of list elements, the method comprising the steps of receiving the speech input, comparing each list element of the first set with the speech input to obtain a first candidate list of best matching list elements, processing the second set using the first candidate list to obtain a subset of the second set, comparing each list element of the subset of the second set with the speech input to obtain a second candidate list of best matching list elements, and selecting a combination of list elements using the first and the second candidate list.

Claims

1. A computer-implemented speech recognition method for selecting a combination of list elements via a speech input, wherein a first list element of the combination is part of a first set of list elements and a second list element of the combination is part of a second set of list elements, the method comprising:

receiving at a processor the speech input;

comparing within the processor each list element of the first set of list elements with the speech input to obtain a first candidate list of best matching list elements;

processing the second set of list elements using the first candidate list to obtain a subset of the second set of list elements;

comparing each list element of the subset of the second set of list elements with the speech input to obtain a second candidate list of best matching list elements; and

selecting a combination of list elements using the first and the second candidate lists wherein selecting a combination of list elements comprises;

determining combinations of a list element of the first candidate list with a related list element of the second candidate list;

scoring each determined combination by combining the score of the list element of the first candidate list and the score of the related list element of the second candidate list; and

determining a result list wherein the result list comprises best matching combinations of a list element of the first set of list elements and a list element of the second set of list elements.

2. The method according to claim 1, wherein one or more list elements of the second set of list elements are related to at least one list element of the first set of list elements.

3. The method according to claim 2, wherein processing the second set comprises selecting only list elements from the second set which are related to one or more list elements of the first candidate list.

4. The method according to claim 1, wherein the first set and the second set correspond to different levels of a hierarchy of sets.

5. The method according to claim 1, wherein the first candidate list is a subset of the first set or wherein the second candidate list is a subset of the second set.

6. The method according to claim 1, wherein the first set and the second set correspond to different categories, in particular, wherein each list element of the first set corresponds to a first category and each list element of the second set corresponds to a second category.

7. The method according to claim 1, wherein comparing the list elements of the first set and/or the list elements of the subset of the second set with the speech input comprises determining for each list element of the first set and/or of the subset of the second set a score.

8. The method according to claim 1, wherein for comparing the list elements of the first set and the list elements of the subset of the second set with the speech input, each list element of the first set is extended by at least one wildcard symbol.

9. The method according to claim 1, wherein scoring each determined combination comprises adding the score of the list element of the first candidate list and the score of the related list element of the second candidate list.

10. The method according to claim 1, wherein determining a result list is followed by comparing each combination of the result list with the speech input thereby obtaining a pruned result list wherein comparing each combination with the speech input comprises determining for each combination a score.

11. The method according to claim 1, wherein the step of receiving the speech input is followed by providing the first and the second set of list elements, wherein providing the first and the second set comprises selecting the first and the second set from a database based on the speech input.

12. The method according to claim 1, wherein the first candidate list, the second candidate list, the result list and the pruned result list comprise a maximum number of list elements, in particular wherein the maximum number of list elements of the first candidate list, the second candidate list, the result list and the pruned result list are determined based on the length of the speech input.

13. A computer program product having computer code on a tangible computer-readable storage medium for selecting a combination of list elements via a speech input, wherein a first list element of the combination is part of a first set of list elements and a second list element of the combination is part of a second set of list elements, the computer code comprising:

computer code for receiving at a processor the speech input;

computer code for comparing within the processor each list element of the first set of list elements with the speech input to obtain a first candidate list of best matching list elements;

computer code for processing the second set of list elements using the first candidate list to obtain a subset of the second set of list elements;

computer code for comparing each list element of the subset of the second set of list elements with the speech input to obtain a second candidate list of best matching list elements; and

computer code for selecting a combination of list elements using the first and the second candidate lists;

wherein the computer code for selecting a combination of list elements comprises;

computer code for determining combinations of a list element of the first candidate list with a related list element of the second candidate list;

computer code for scoring each determined combination by combining the score of the list element of the first candidate list and the score of the related list element of the second candidate list; and

computer code for determining a result list wherein the result list comprises best matching combinations of a list element of the first set of list elements and a list element of the second set of list elements.

14. The computer program product according to claim 13, wherein one or more list elements of the second set of list elements are related to at least one list element of the first set of list elements.

15. The computer program product according to claim 14, wherein processing the second set comprises selecting only list elements from the second set which are related to one or more list elements of the first candidate list.

16. The computer program product according to claim 13, wherein the first set and the second set correspond to different levels of a hierarchy of sets.

17. The computer program product according to claim 13, wherein the first candidate list is a subset of the first set or wherein the second candidate list is a subset of the second set.

18. The computer program product according to claim 13, wherein the first set and the second set correspond to different categories wherein each list element of the first set corresponds to a first category and each list element of the second set corresponds to a second category.

19. The computer program product according to claim 13, wherein the computer code for comparing the list elements of the first set comprises computer code for determining for each list element of the first set a score.

20. The computer program product according to claim 13, wherein the computer code for comparing the list elements of the first set and the list elements of the subset of the second set with the speech input, each list element of the first set is extended by at least one wildcard symbol.

21. The computer program product according to claim 13, wherein the computer code for scoring each determined combination comprises computer code for adding the score of the list element of the first candidate list and the score of the related list element of the second candidate list.

22. The computer program product according to claim 13, wherein determining a result list is followed by comparing each combination of the result list with the speech input thereby obtaining a pruned result list wherein comparing each combination with the speech input comprises determining for each combination a score.

23. The computer program product according to claim 13, wherein the computer code for receiving the speech input is followed by computer code for providing the first and the second set of list elements, wherein providing the first and the second set comprises computer code for selecting the first and the second set from a database based on the speech input.

24. The computer program product according to claim 13, wherein the first candidate list, the second candidate list, the result list and the pruned result list comprise a maximum number of list elements, wherein the maximum number of list elements of the first candidate list, the second candidate list, the result list and the pruned result list are determined based on the length of the speech input.

25. An apparatus for speech recognition for automatically selecting a combination of list elements via a speech input, wherein a first list element of the combination is part of a first set of list elements and a second list element of the combination is part of a second set of list elements, the apparatus comprising:

a receiving module for receiving the speech input;

a first comparison module for comparing list elements of the first set of list elements with the speech input to form a first candidate list;

a processing module for processing the second set of list elements using the first candidate list to produce a subset of the second set of list elements;

a second comparison module for comparing list elements of the subset of the second set with the speech input to form a second candidate list; and

a selection module for selecting a combination of list elements wherein selecting a combination of list elements requires;

determining combinations of a list element of the first candidate list with a related list element of the second candidate list;

scoring each determined combination by combining the score of the list element of the first candidate list and the score of the related list element of the second candidate list; and

determining a result list wherein the result list comprises best matching combinations of a list element of the first set of list elements and a list element of the second set of list elements.